How Spotify's new advertising strategy impacts artists

I. The artist is the advertiser

“Where is the market, and what is actually being sold?”

A group of Swedish researchers asked this question in their 2019 book Spotify Teardown: Inside the Black Box of Streaming Music — a critical history of Spotify that the company itself didn’t want published. The context of this question was a simple but still illuminating discussion of the foundations of Spotify’s advertising business.

To this day, the majority of Spotify’s users (210 million, or 57% of the total) do not pay to use the service. The question that follows is: For this majority user base, what is Spotify actually “selling”? As with so many free media platforms today, the users themselves become the product for advertisers. “The real ‘value of free’ is that it helped scale Spotify’s marketplaces” to the level where it would be attractive to outside brands, with music being “promoted as merely functional for defining and micro-targeting audience segments,” wrote the Spotify Teardown authors. From the very start, “[Spotify] was unequivocally in the business of providing content to audiences while selling those audiences to advertisers.”

This argument might initially sound strange; after all, Spotify has not historically framed themselves to the music industry as an “advertising-first” platform. Rather, their branding throughout the early to mid-2010s focused more on automated, personalized discovery of new music for fans; democratized, global, anti-gatekeeper distribution for artists; and a fairer, more lucrative alternative to piracy for the industry as a whole.

That all changed in 2018 and 2019, when Spotify went public and began trying to offset its licensing costs by investing aggressively in the podcast industry, which still makes the majority of its revenue from ads. Today, the way Spotify frames itself to investors and audio-industry stakeholders is absolutely as an advertising- rather than music-first business. Their latest earnings release touted a 627% jump year-over-year in podcast ad revenue, to the point where CEO Daniel Ek claimed that “the days of our ad business accounting for less than 10% of our total revenue are behind us.” Many of Spotify’s recent news announcements, from eight-figure exclusive podcast deals to the launch of the Spotify Audience Network, point to confidence in their on-platform ad growth. Even Dawn Ostroff’s title at the company — “Chief Content & Advertising Business Officer” — illustrates how Spotify sees content and ads growing hand-in-hand moving forward.

Many industry critics will point to how Spotify has built their multibillion-dollar valuation off the backs of artists’ catalogs, but artists themselves seem not to be reaping the rewards of this growth. In the specific context of advertising, as the authors of Spotify Teardown put it, “the benefits for musicians … whose content is the incentive attracting both sides” (i.e. the subscriber and the advertiser) remains unclear.

But this critique also makes an inherent assumption about Spotify’s strategy that is now false — namely, that artists and advertisers are two completely separate stakeholders in the Spotify ecosystem.

Up to this point, it has certainly felt that way. Spotify Advertising and Spotify for Artists are two totally separate landing pages and PR initiatives; the former caters to helping larger companies and brands with bigger marketing budgets to spend, while the latter seems to focus on individual, emerging artists who might be more strapped for cash but still want to build a sustainable career, business and fan base.

As you’re reading this, a major shift is happening in Spotify’s strategy in which these two previously separate groups of stakeholders are now merging. Namely, in Spotify’s eyes, the artist is now one and the same with the advertiser. And if they want the privilege of reaching the right listeners on Spotify’s platform, they will have to pay up.

II. Commercial Considerations Rule Everything Around Me

Nearly all of the features Spotify has been pushing to the music industry in recent months operate under this assumption of treating the artist as the advertiser.

Three tools stand out in particular:

A. Ad Studio — A tool that allows artists, or any brand for that matter, to create and launch audio ad campaigns to promote their songs. These campaigns reach free users only, and can involve multiple forms of targeting including location, genre, interest (e.g. books, comedy, fitness, studying), “real-time context” (e.g. moods and activities) and fan base (targeting your existing fans, or those of similar artists). Ad Studio emerged out of beta at the start of the COVID-19 pandemic, and is a fairly straightforward tool that has not gotten much critique — likely because it’s similar to what many other freemium streaming services like Pandora and JioSaavn have offered for years.

B. Discovery Mode — A tool that allows artists and labels to pick specific songs from their catalogs to prioritize in Spotify’s Autoplay and Radio features, for both free and paying users. While there is no upfront budget to use the tool, streams for songs in Discovery Mode receive a lower royalty rate in return for extra promotion.

The independent music sector’s reaction to Discovery Mode has been surprisingly polarized. Several self-serve distributors like DistroKid and TuneCore have praised Discovery Mode as “highly innovative and beneficial for independent artists,” while indie trade and advocacy groups like A2IM, IMPALA and the Artist Rights Alliance — and even government organizations like the U.S. House Judiciary Committee — have directly or indirectly equated Discovery Mode with radio payola.

In any case, one can’t help but think that the framing of “Discovery Mode” is potentially misleading; “discovery” implies some level of natural spontaneity or hands-on serendipity, whereas this particular ad tool was designed with the intention for strings to be pulled — or, in this case, pushed — behind the scenes.

C. Marquee — A paid advertising tool that allows artist and label teams to deliver fullscreen notifications about new releases to specific segments of their audience base.

Unlike with Discovery Mode, Spotify makes Marquee’s upfront costs plain and simple, paving a much clearer path for evidence-based critique from the music industry. According to an official FAQ page, artists can set up campaigns for up to 21 days after a new song or album is released on the platform, with budgets ranging from $250 to $10,000. The artist gets charged on a per-click basis — $0.50/click for targeting one’s general “reachable audience,” and $0.55/click for targeting specific audience segments like “recently interested listeners,” “casual listeners” or “lapsed listeners.” Campaigns run for up to a week or when artists go through their budget, whichever comes first.

On August 6, Fourth Strike Records — an indie label focused on LGBTQ+ artists across punk, experimental and other genres — posted a scathing, viral Twitter thread about Marquee, with a focus on why the economics of the tool don’t make sense for their roster. Assuming an average per-stream payout rate of $0.004 (this figure varies among different artists and DSPs), Fourth Strike did their own back-of-the-napkin math: “to make our money back per-tap, we’d have to get 125 streams — and that’s not accounting for transaction fees, distribution fees, etc.,” they wrote in their thread. “each person who tapped this screen would have to listen to the album front-to-back 12 times.”

For context, the band with the largest streaming audience on Fourth Strike Records’ roster, The Garages, has only around 7,000 monthly listeners on Spotify. Looking at the high levels of fan engagement on the label’s Bandcamp pages, my educated guess is that they’re not making most of their money from streaming, anyway. Still, the economics do feel a bit off, especially if you look at Marquee’s fine print, which states that an artist needs a minimum of 15,000 streams in the U.S. over the last 28 days in order to qualify for Marquee. Assuming the same $0.004 per-stream rate, that means a streaming income of only $15 per week — and yet Marquee is asking for a minimum campaign spend of $250 per week.

More generally, tools like Marquee run the risk of eroding the brand equity that Spotify has built over the past decade with respect to making discovery more democratized and accessible for artists. “Spotify charges you to listen to music, and now they want to charge artists for the *chance* to be listened to, even by their own fans,” Fourth Strike wrote on Twitter. “They strip the only selling points they have, their supposed ‘discoverability’, in order to charge artists for them. it’s absolutely fucked.”

Spotify has been experimenting with ways for labels to pay to promote songs on their platform for at least four years — and not just through audio ads. In 2017, Josh Constine found for TechCrunch that Spotify was testing “Sponsored Songs” that labels could place at the top of official and user-generated playlists. Marquee itself has been around since 2019, at which point it was more targeted at bigger artists and labels, with a recommended minimum budget of $5,000 per campaign.

What’s driving more awareness and critique around Marquee is the fact that it’s rolling out to a much wider range of artists with a much lower floor of activity, even if it might not make financial sense for them. Plus, the proliferation of all of these ad tools may be directly impacting the integrity of the listener experience on Spotify. As music/tech writer Bas Grasmayer pointed out in the Water & Music Discord server, Spotify now features an “About recommendations” option in the menus for all of its official playlists (both editorial and algorithmic), with the following disclaimer hidden at the bottom:

In some cases, commercial considerations may influence our recommendations, but listener satisfaction is our priority and we only ever recommend content we think you’ll want to hear.

If you connect the dots between this disclaimer and the fact that Spotify Marquee and Discovery Mode advertising campaigns touch both free and paying users, you come to the conclusion that the days of Spotify Premium being “ad-free” are officially behind us. Sure, as a paying subscriber, you might not get any “interruptive” audio ads while you listen. But there are still signs of paid advertising baked across the Premium experience — from dynamic ad insertion in podcasts, to the Announcements banner on the homepage, to now fullscreen Spotify Marquee notifications and intentionally pushed songs in Discovery Mode.

Just because you’re a Premium Spotify subscriber doesn’t mean you won’t still be subject to the immense competition among artists to reach your ears. And as an artist on Spotify, you’ll now be subject to the platform’s commercial incentive to gate-keep not only discovery, but also attention, and make you a paying part of that competition.

III. Attribution, ROI and other fuzzy terms

Let’s home in on Spotify Marquee — specifically on the way that Fourth Strike Records frames the tool’s economics, and what it reveals about the highly ambiguous nature of digital music marketing today. There are two main concepts to focus on here: Attribution and return on investment (ROI).

In their Twitter thread, Fourth Strike rightfully points out that Marquee requires the artist “to pay upfront for their own followers to see that they’ve released a new single or album,” as illustrated by their following screenshot:

What this screenshot does not reveal, however, is that Marquee also allows the artist or label to target three additional sub-segments of their reachable audience (for an additional 5¢ per click):

• Recently interested listeners — those who “actively streamed your music in the last 28 days,” in Spotify’s words.
• Casual listeners — those who “actively streamed your music in the last 6 months, but less than your other listeners.”
• Lapsed listeners — those who have some listening history with your music, but “haven’t actively streamed your music in the last 6 months.”

Within each of these segments, artist teams are able to measure not only the conversion rate on a Marquee campaign (% of listeners who “streamed at least one track from the release within 14 days after seeing your Marquee”), but also the intent rate (% of listeners who “saved or added a track from the release to a playlist”).

From a marketer’s standpoint, the honest reality is that this level of segmentation and attribution for music streaming campaigns is better than literally any other alternative on the market right now — at least for Spotify.

A bit more context: Attribution modeling — or the act of determining which specific channels, advertisements or messages have the greatest impact on a customer’s decision to engage with a given product — remains one of the biggest challenges in modern music marketing. Let’s illustrate this with a hypothetical scenario. Say you’re Fourth Strike Records, and you start a paid advertising campaign on Facebook around your new album. Through the ad, you drive viewers to a smart link (using a tool like Linktree, Linkfire or Feature.fm), which then enables them to listen to the album on the streaming service of their choice.

All in all, say the campaign ends up driving a 200% spike in streams on Spotify. But figuring out who exactly drove that spike and how exactly they engaged with the album is much more difficult to pin down. In terms of data, you can get surface-level information on how many people clicked the Facebook ad, and then how many people clicked through to listen on Spotify. But because Facebook and Spotify silo their raw user data from each other and from third parties (for legal and competitive reasons), there’s no way to paint a complete picture of the fan journey from the first ad impression across their engagement with the album on Spotify over time, let alone to determine how that journey looks different across various listener segments.

Back in August 2020, artist and software engineer Kimberly Hou wrote for Water & Music about the lack of attribution and referral data within Spotify for Artists’ own dashboard. Hou envisioned a “Google Analytics for music streaming,” with the ability to measure clicks, referrals and overall lift within a custom date range, but found that Spotify for Artists’ data dashboards fell far behind Silicon Valley standards. For instance, “seeing a spike of recent streams coming from my Spotify profile instead of from listeners’ own playlists might indicate that new listeners are discovering my music from elsewhere,” wrote Hou, but the limited amount of data she can see on her end means that it’s difficult to craft an “improved outreach strategy outside of Spotify … Did this spike in listeners on my profile come from a social media post or a press article? If it came from a group of people who shared my music online, whom can I thank directly?”

Without clear attribution, marketers run the risk of spending their digital budgets blindly. Third-party smart-link startups like Linkfire and Feature.fm do have some attribution data for DSPs like Apple Music, Deezer, Pandora, Anghami, Boomplay and Qobuz — but not for Spotify. And even these startups’ attribution data focuses more on conversion to surface-level streams, without the ability to distill differences in listening/playlisting activity across specific segments. Marketers need access to this latter data to determine, for instance, whether a campaign successfully converted casual or lapsed followers into diehard, loyal devotees.

With this background in mind, Spotify’s public case studies about Marquee with artists like Mt. Joy and girl in red emphasize the ability to measure conversion and intent across specific segments, in a way that wasn’t previously possible. Of course, Spotify still ultimately owns all of this listener data, and is still delivering these insights on their terms and within a closed product ecosystem, without the ability to connect the data to social platforms like Facebook, Instagram or Twitter where the music industry is already actively spending its marketing dollars.

It’s also unclear what the payback periods for Spotify’s ad tools look like — i.e. how much time it takes the average artist to make their Marquee budget back, or how many streams it takes to make up for the lower Discovery Mode royalty rate, and whether that money spent is worth it compared to traditional social media ad campaigns. This leads us to our second fuzzy topic of ROI.

In their Twitter thread, Fourth Strike framed their ROI on a Marquee campaign in terms of how many times each listener would have to stream the specific album being promoted, in order for the label to make their money back per-click. To recap the math: 5¢/click divided by $0.004/stream = 125 incremental streams per click to break even. Again, for an indie label whose artists have only a few hundred to a few thousand monthly listeners each, achieving ROI on this kind of campaign will likely be difficult and/or take a disproportionately long time.

With more direct targeting capabilities, Marquee will hopefully have a shorter payback period than your traditional audio ad campaign. But at the same time, measuring ROI on a digital music ad campaign might not be as simple as looking at the return on just one specific catalog or album. What if a lapsed fan clicked your Marquee ad and streamed your songs just a few times — but then got inspired and went on your website to buy a bunch of your merch? On an individual level, the ROI on that Marquee campaign with respect to streams is low, but the ROI on that fan as a whole looks much better when you look across income sources.

In TV advertising, marketers often refer to this dynamic as the “halo effect” — where the return on TV commercials comes not from immediately convincing viewers to buy a given product at that moment, but rather from cultivating more positive long-term awareness so that viewers have a higher chance of remembering and paying for those products down the line.

Similarly, the ROI for paid music streaming campaigns like those launched through Marquee will likely come more from maximizing longer-term audience development and fans’ lifetime value both on and off of Spotify, rather than just from incremental streams of a specific song or album in one week. But it’s unclear whether Spotify is interested in putting the benefits of Marquee in this long-term context, given the company’s incentives to promote the value of its own, closed marketing ecosystem.

IV. Advertising as platform power

In short, Spotify treating the artist as the advertiser might be good news for music marketing, especially with respect to better audience segmentation and attribution. But it also charts a potential future path for Spotify and its artist relationships that’s not-so-good news with respect to platform power.

With this in mind, I want to close out this piece with a discussion of two articles that contextualize why Spotify’s new approach to advertising is so polarizing in the music industry, especially among independent artists and labels.

The first resource, h/t Future of Music Coalition, is a paper that Lina Khan published in the Georgetown Law Technology Review a few years before her appointment as the latest Chair of the U.S. Federal Trade Commission. It’s called “Sources of Tech Platform Power,” and outlines three sources of power in particular that big-tech companies wield today:

• Gatekeeper power — where platforms that dominate as the primary intermediary infrastructure for digital markets can “extort and extract better terms from the business users that depend on their infrastructure,” in Khan’s words.
• Leveraging power — where platforms leverage their dominance “to establish an advantageous position in a separate or ancillary market.”
• Information exploitation power — where platforms use data and “insights generated by third-party retailers and producers to go head-to-head with them, rolling out replica products that it can rank higher in search results or price below-cost.”

Multiple aspects of Spotify’s recent growth — from Spotify Marquee and other ad tools, to their overall dominance of the podcast market — involve all three kinds of platform power:

• Gatekeeping: Spotify is not so much a music discovery engine as it is a music discovery broker, using the phrase of a “two-sided marketplace” to try to extract more value out of a music industry that is desperate to reach the right audience. (If the world of social media is any indication, it’s likely that listener/fan acquisition costs will also rise on Spotify as their ad tools become more popular.)
• Leverage: Spotify has been using its success in music streaming to try to assert its dominance in podcasting, and subsequently in audio advertising.
• Information exploitation: This applies to both content and ads. By analyzing data around third-party podcast content, Spotify can optimize production and then prioritize marketing for its own original podcasts (similar to Netflix). And with tools like Marquee and Discovery Mode, Spotify is trying to use its dominance of music listening data to build internal ad tools for the music industry that arguably compete directly with third-party platforms like Linkfire and feature.fm.

The second resource is investor Li Jin’s latest newsletter, which focuses on how supposedly “creator-focused” tech companies are at risk of recreating the exact platform inequality and worker risk that the creator economy was supposed to fix. With an enlightening turn of phrase, Jin calls centralized tech platforms like Facebook and Instagram not democratized access points between creators and fans, but rather “access chokepoints.”

Her reasoning is that these platforms still “own the data, social graphs, and end user relationships — all of which creators need in order to access audiences and income … in the majority of cases, this type of capital cannot be easily ported over to external, creator-owned properties. In this way, creator labor is controlled and commoditized by platforms.” This dynamic also has a direct line to how advertising dominates social media today: The commoditization of creator labor allows tech platforms to “maintain their grasp on user attention, a necessary ingredient for advertising-based business models.”

The effect of Spotify treating the artist as the advertiser has a similar feedback loop: If tools like Marquee and Discovery Mode work, then listeners will pay more attention to artists’ catalogs and spend more time on Spotify, therefore becoming more valuable to more artists and brands who want to pay to advertise on Spotify’s podcasts and overall platform. The artist simply becomes a commoditized cog in this flywheel.

To revisit Spotify Teardown’s question: “Where is the market, and what is actually being sold?” By treating the artist as an advertiser, Spotify is morphing from a music market to an attention market, from right under millions of unsuspecting buyers’ noses.